it's so funny how much rumour circulates around this issue... i posted a 
link in this thread to the corresponding manual section which the op 
might have chosen to ignore. oh well...


> to be honest, I have never understood the rationale why the source version
> should not be used with a web server

the rational is very simple, and it's *not technical*: running the 
source version from disk is what we *support*. it's in our programming 
model. that doesn't mean you cannot do it differently, but we cannot 
support each and every possible scenario. we try to give hints, but we 
don't fight for it. if you get it running otherwise - fine. if we change 
something that works in the from-disk scenario, but breaks yours, your 
on your own. it's that simple.

>  - I have never had any problems with
> this setup,

it's not difficult, if you get the key issue, like Derrell put it, and 
know what to do on your file system and web server config. but this is 
not generally true.

>  as long as you put the CACHE into the source tree itself. 
>   

now, this line on it's own is nearly wrong, and probably adds as much 
confusion as it might provide help.

the truth is, slightly re-phrasing what Derrell wrote: all libraries of 
your qooxdoo application (your app is a library, the framework is a 
library, any contrib is a library,...) must be reachable from the web 
server, and relative paths between them on the file system must result 
in valid URL paths when served through the web server. with apache, 
there are several ways of achieving this (one of which is using aliases, 
which hasn't been mentioned so far).

in the case of contrib libs referenced with the "contrib://" pseudo 
protocol, that means your download path has to be reachable from your 
web server in the above sense, and pointing the CACHE config macro into 
your source tree is a *very coarse* approach to achieve this. the 
download path can just as well be the default location 
(<system-tmp-dir>/cache/downloads), but can be any other path if you 
adapt the config key "cache/downloads" for your build jobs accordingly. 
the *compile* cache has nothing to do with it, so there is no need to 
shuffle it around (which you do when you change CACHE).


t.



------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
_______________________________________________
qooxdoo-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/qooxdoo-devel

Reply via email to